In a variety of clinical situations it is important to measure certain chemical characteristics of the patient's blood such as pH, concentrations of calcium, potassium ions and hematocrit, the partial pressure of O.sub.2 and CO.sub.2 and the like. (See, for example, Fundamentals of Clinical Chemistry, Tietz, Editor, page 135 et seq., Electrochemistry; page 849 et seq., Blood Gases and Electrolytes; 1976 Saunders Phila.; see also the patent to Battaglia et al. U.S. Pat. No. 4,214,968.) These situations range from a routine of a patient in physician's office to monitoring during open-heart surgery. The required speed, accuracy, and similiar performance characteristics vary with each situation.
Measurement of chemical characteristics of blood during open-heart surgery provides the most demanding set of criteria. Presently, blood gas analysis during major surgery is provided by repeated transfer of discrete blood samples to a permanent lab-based blood gas analyzer or by use of sensors placed inline with the extra-corporeal blood circuit of a heart-lung machine employed to bypass the patient's heart.
The transfer of discrete blood samples, required by blood-gas analyzers inherently increases the risk of contaminating the blood sample with ambient air, which may alter certain of the monitored characteristics. Additionally, since such analyzers are complex and costly devices, they are typically located only in the hospital lab where they need to be operated by a skilled technician, resulting in undesirable delay during surgery, critical care or intensive care. Further, such analyzers employ bubble tonometers to generate a suitable gas reference mixture by dissolving quantities of gases, stored in pressurized free-standing tanks, into the electrolyte solution. While replacement of these gas tanks is infrequently required, it is a cumbersome procedure. Finally, these existing analyzers require cleaning to decontaminate all exposed portions from the prior patient's blood prior to subsequent use.
Although use of in-line sensors minimizes both the risk of contamination during transfer and the risk of delay, these sensors have a response which normally varies or "drifts" during use; moreover, this drift is not at a constant rate. Present in-line sensors can only be calibrated before they are placed in the extracorporeal circuit. Thus, the inherent drift of these in-line sensors cannot be monitored, resulting in readings of ever decreasing reliability as time passes.